Tuesday, June 15, 2021

Again: Ability Scores

I've been spending a lot of time working out the details of The Vaults of sha-Arthan setting I hope to start refereeing in July or August. While I'm using Old School Essentials as the base for its rules, I'm also making changes to many aspects of those rules, such as the character classes, magic, monsters, treasure – well, quite a lot, now that I think about it. Even so, it's my intention that sha-Arthan be recognizable as a descendent of Dungeons & Dragons, much in the way that Empire of the Petal Throne is.

At the moment, I'm devoting my time to writing the rules that players need to generate characters, including the matter of ability scores. As you may recall from my recent posts about ability scores, I have a number of questions and criticisms of the traditional D&D ability scores. At the same time, ability scores are such a foundational element of D&D – and indeed pretty much all traditional RPGs – that I wasn't all that keen to dispense with them entirely, though I did seriously consider it.

My broad thoughts are the following:

  1. In principle, I very much like the idea of randomly generated ability scores, preferably on the 3d6 in order model. 
  2. One of the reasons I like random generation is that it offers a good chance that each character has at least one below average score, which I think lends just as much uniqueness to character as the abilities where he is above average.
  3. That said, if ability scores provide significant mechanical benefits, my fondness for the idea diminishes somewhat.
  4. While I am very much in favor of limiting the presence of certain "special" character classes and races, I am quite convinced that doing so by recourse to ability score minimums only feeds the perceived need for characters to have higher scores and for the generation systems to support that need.
  5. Finally, I increasingly think there ought to be some system by which a character can improve his ability scores over time, whether it be through training as in RuneQuest or gaining levels as in Empire of the Petal Throne or post-TSR versions of D&D.
With all that in mind, I'm now weighing two options for generating ability scores in light of these thoughts. The first and most straightforward is also the most radical: dispense with random rolls. If I take this approach, there seem to be two options: point buy or a typical distribution. In the former, I'd settle on some number of points, say 70, and then allow each player to divide those points among his character's ability scores as he wishes (with 3 being the minimum and 18 the maximum, of course). In the latter, each player distributes the same collection of scores, say 6, 8, 10, 11, 11, 12, 14, and 16 amongst his character's abilities. The advantage of these two approaches is that, while they entirely eliminate the random aspect, they greatly increase – or ensure, in the case of the typical distribution – that a character will have below average scores.

The second approach preserves a bit more of the random element I like. I came across a version of it in the second edition of The Black Hack. In the system presented there, ability scores are rolled 3d6 in order. However, if the score for any ability is 14 or higher, the next ability score is not rolled but is rather simply assigned a 7. The Black Hack also permits the player, after he has rolled all six scores, to swap two of them in order to better suit his intentions for his character. I find this approach strangely attractive, since it preserves the random element while placing the proverbial finger on the scale in order to ensure any high score comes at a cost.

There are probably other alternative approaches to generating ability scores that address my concerns and, if so, I'd love to know about them. For now, I continue to weigh the advantages and disadvantages of the approaches presented above, without having made a firm decision toward one or the other. Right now, what's most important (to me anyway) is that I have made my peace with the idea of not using 3d6 in order, something I still very much prefer in principle but that I realize I had been turning into something of an idol. I now have a much better sense of what I want out of ability scores in sha-Arthan; all that remains is deciding the best way to get what I want.

43 comments:

  1. Hmm, in Black Hack is there a corresponding "if you roll a 7, the next score is not rolled but is assigned 14"? Also what order are the attributes in? That scheme mostly assures a character who is for example (assuming the order of attributes is STR, DEX, ...) strong is low in dexterity (if the order is STR, DEX, ...)

    I've gone all over the place and in the end, I think I prefer some form of roll attributes in order allowing dropping of a really bad set of rolls. I don't want to preclude the character who is just awesome all around. Yea, that means that character outshines the other, but the best way of assuring that one character doesn't dominate the game is not mechanically limiting the character, but rather making enough stuff important that no one character can do it all.

    Even my RQ campaign with a near perfect elf, there are things the other PCs do better than the elf. And I've learned my lesson, no more elves... (sounds like a Talislanta add :-) )

    ReplyDelete
    Replies
    1. What you point out about The Black Hack is something I'd thought about too. I think the reason you can swap two scores is to avoid all the really strong characters being stupid, for example, but that's probably not enough to overcome the inevitable sameness that its approach would engender.

      Delete
  2. Stats seem so important if you assume that one player only ever has one character at a time. I used to tie myself into knots looking for the best way to roll stats.

    When I first heard folks like Tim Kask talking about how back in the 70s, they always had five or six characters at a time that they were "working up" in the campaign — a stable of choices to draw from when deciding who to take on a given adventure — that was a Eureka moment. The scales fell from my eyes, and 3d6-in-order without modification (among many other wonky-looking rules) suddenly made sense.

    ReplyDelete
    Replies
    1. That certainly makes sense and I'm grateful that you mentioned it. However, I've never played in a campaign where the players had more than a couple of PCs at any given time, so I'm not entirely sure this approach would work for me. It does offer some useful perspective, though.

      Delete
    2. Of course, that was also back when, for the most part, ability scores meant little more than potential modifiers to XP gained.

      Once the ability score modifiers grew to become serious game changers, the importance of ability scores grew to, in some cases, being obsessive. Especially with AD&D, when bonuses were reserved for the highest scores.

      That's why I like Moldvay modifiers; it spreads them out a bit. +1/-1 is commonplace, +2/-2 is not unusual, but +3/-3 is rare, and folks usually have enough modifiers to go around... plus, of course, in Moldvay, you were also allowed to exchange points.

      Delete
    3. The stable of characters is certainly an alternative. The problem is that unless you at least play every week for a significant session, and ideally multiple times a week, switching between multiple characters is going to mean not very much play time for each character.

      One way to get some of the effect though is to roll multiple characters, and play your favorite.

      But it's still possible to have horrible luck and roll a string of poor characters.

      Delete
    4. I never really did character stables until Dark Sun came along, but since then I've become something of a fan of the idea, as well as Ars Magica troupe-style play where there are "extras" (grogs, in AM) that anyone and everyone can play. I've often thought that might be a good approach to truly old school delving, with most players in any given adventure standing in the shoes of the hirelings and henchmen of one or two "real" PCs.

      Delete
    5. Eh, rolling multiple times and picking what you want is not really that different from a balanced point assign. I knew some gaming groups who rolled as many as 10 characters, picking the best one and proudly claimed that they did not do a point buy. They were practically almost fooling themselves.

      Delete
    6. @Dead-meat You're not wrong, that's for sure. I can remember folks in AD&D who had whole notebook pages covered in stat arrays they'd rolled up during study hall. Pointless exercise IMO.

      Delete
    7. Rolling until you get what you want is not point buy, or at least not balanced point buy. To some extent it's even random as long as you are only seeking a small portion of the attributes. The "hunt" and "service" options I added to the Classic Traveller Online Character Generator (https://ffilz.github.io/Gaming/travellercharacter.html) allow specification of a few elements of a character but otherwise you wind up with a random character.

      But I think "roll until you get what you want" is very different from "roll 2-4 character and pick one". And I really don't see either as point buy.

      The problem I have with point buy is that in some systems you never see the maximum attribute (because it costs so much for what you get) and the tradeoffs really push you towards a relatively average character. It also highlights dump stats. It also may not work well when one character concept requires 2-3 good attributes while another only requires 1 good attribute.

      Delete
    8. I see what you mean. For what it's worth, what I meant to say was that if you roll as many times as you need to get what you want, the result is no different than simply assigning the points you want in the first place. You just waste more time. Well, some people enjoy the rolling, so I can respect that.

      Delete
    9. The Funnel method (I believe due to Goodman Games) is another version of this: roll up a number of level zero characters per player, run them through an introductory adventure (perhaps all at once), and those that survive and gain enough XP advance to level one and can become PCs, making a stable that’s been initiated en masse.

      I’ll probably use this when I start an EPT campaign (modified by Heroic Age of Tékumel rules) with the Fresh Off The Boat premise. Start the characters as non-adventuring “barbarians” who get driven off their island by some catastrophe; those that survive get enough training for level one on the long trip to Jakálla.

      Delete
    10. DCC's funnel is fine for the system. Quick and easy.

      Delete
    11. While DCC may have introduced the stable of characters funnel, the idea of starting PCs at level 0 was introduced in TSR Module N4 Treasure Hunt (at least that's where I first saw the idea). Level 1 character class was determined by the PC's actions in the adventure.

      Delete
    12. Oh, yea, rolling until you get what you want is similar to point buy, except that each player is picking their own point score, and justifying it because the "rolled it fair and square". Rolling a fixed number of times is definitely different, or rolling until some threshold that applies to all (like roll until you get at least one 16 or higher).

      Delete
  3. On one hand, I want to denounce your heresy. On the other hand, I think that whatever you do will turn out great, so I'm curious. Very curious.

    ReplyDelete
  4. I thought I'd add a separate comment on the method I used for one RuneQuest campaign. Players got a number of points to distribute among their attributes (108?). But instead of calculating the ability bonuses from those attributes, players got a standard array of ability bonuses (+10, +15, +15, +20, +20, +25) to distribute across the set of ability bonuses. Each attribute still had it's own impact and was worth having a good score in, but it was no longer so critical.

    But it's not random... So my current campaign is using 4d6k3 for humans (with INT being 3d6k2+6 to reflect the later suggestion that INT be 2d6+6). We seem to have decent PCs. We do still have ONE PC rolled on straight 3d6 (2d6+6 INT) from the very start of the campaign.

    ReplyDelete
  5. I like the Stars Without Number system where you either choose a standard array (14, 12, 11, 10, 9, 7) and arrange as you like *or* roll 3d6 in order but at the end you can change one score to 14. I've got some players who hate to gamble and just want to be sure they get a viable character for their concept and others who are all about roll the dice to see what happens. SWN, though, doesn't make the stats all that important in the end, so nobody really outshines another player because they've got a +1 instead of a +0. A system that used a lot of roll a d20 under your stat to succeed makes having low scores pretty punishing.

    I've also had some fun with a variant where you roll 3d6, but you assign as you go; once assigned you can't swap. That lets players steer their characters towards a concept, but not really optimize.

    ReplyDelete
    Replies
    1. Stars Without Number is a very good game. That needs to be said more often.

      Delete
    2. James, if you have any interest in possibly making a conversion to BRP, I'd take a look at the "Ware Hall" solo adventure in Different World's # 17. That's probably the most rules lite version of the D100 rule system I've come across yet.

      Delete
    3. "Ware Hall" looked like fun and I considered actually playing it. I have a soft spot for solo adventures, so I might just do that in the near future.

      Delete
  6. Have you thought of having PCs generated by rolling 1d6+7 for each stat?
    This way you can keep modifiers low, have everybody more or less within a small range, and it leaves you ample room for growth (since you are thinking of giving characters the chance to increase ability scores...).
    If you wanted the PCs to have a wider range of scores you could go with 2d6+3, but I think 1d6+7 might work better

    ReplyDelete
  7. Doesn't 5E have ability score increases as a level-up perk every once in awhile?

    ReplyDelete
    Replies
    1. You can swap a Feat for an ability score increase in 5e. 3.x and 4e have once-in-a-while increases

      Delete
    2. Don't some 5e feats also come with a stat bump baked into them as well? My memories of the rule set are surprisingly hazy, but I've played fewer games of that edition than any other - at least past the playtest period when our 4e group was trying to choke down the changes. That group gave up on 5e immediately after 13th Age released, to my great relief.

      Delete
    3. Yes, some Feats come with a stat bonus.
      I do like 13th Age more than 5e as well, btw.

      Delete
    4. I confess I do rather miss the "board game" style combat that 3.0/3.5 and 4e had gotten me used to, but 13th Age has so many other things I really enjoy that I'll cope with going back to Theater of the Mind fights. Faster, too - although some of that's the Escalation Die mechanic, which I've taken to using in other systems as well.

      Delete
  8. Everyone brings up 3d6 in order, but what order? D&D switched from SIWDCC to SDCIWC (A fact I learned last week) with AD&D2. So why not mess with it further?

    As for point buy? I've contemplated making the players buy modifiers on a 1-for-1 basis, so that a +1 in one score will be paid for with a -1 in another score. A +2 with either a -2 or two -1 scores. Simplifies the math.

    ReplyDelete
  9. I'm not entirely sure what the issue is. If the problem is that players game the system to gain a 'winning' character then either change your players or balance the system through the metagame e.g. double the experience required to rise a level for characters with an 18 prerequisite attribute and halve it for those with a 7 or less.

    ReplyDelete
    Replies
    1. The idea of balancing core stats and other elements of your character is an interesting one, but outside of pure point-build-from-the-ground-up systems like Champions you don't see it much. Forbidden Lands does it a tiny bit with older characters having lower stats but more skill points, and several indie RPGs assign you better backgrounds and gear if your stats are awful (and vice versa) - but I can't think of one where character advancement speed is the inverse of their starting strength.

      So, yeah, neat idea. Innovative.

      Delete
    2. The problem I could see with this idea is that your character starts off great because of the high attribute but eventually falls behind and to the extent that scores in things matter (which is why we're even having this discussion) maybe becomes unplayable.

      It's also counter to our general feeling that people who are more suited for X profession from the start tend to stay at the top.

      Delete
    3. @Frank Contrariwise, it would be an excellent iteration of the Peter Principle, wherein the marginally competent are swiftly promoted beyond their level of competence (often by superiors who feel unthreatened by them) while those with native talent stay in the mail room for years. :)

      Delete
    4. True, but the Peter Principle doesn't hold across all endeavors. Also, promotion doesn't equal expertise. And actually, in some fields, you want the best to NOT be promoted, at least not promoted to management.

      Delete
    5. I think of it more along the lines that someone who finds a task easy gains less self-knowledge from success, and thus develops a flexible range of appropriate skills more slowly, than someone who has to struggle.

      Delete
  10. So a mechanical bonus exists, but you don't have it --- so what?

    Magic items exist that your character doesn't have either. Big deal.

    The focus is all wrong. The environment should be defined such that you can succeed without exceptional stats...and if you do get that lucky, they still should only make a small difference. The rest is all prideful nonsense.

    ReplyDelete
    Replies
    1. If you do not care about the character's specific mechanics, then yes. But some like playing characters with specific mechanical traits and stats just for the sake of playing them. It doesn't have to be about winning or the advantage.

      Delete
    2. Sure, looking for something specific is a reason to consider point buy over random generation, though I think there's still room for an in between where you set some goal and roll until you get it. What would be super cool is a system that allowed you to achieve that goal in one shot.

      Also, the more complex the character, the more choice players are likely to want, so a random system that allows for some choice might be desirable (Traveller allows some choice during previous experience for example).

      I am always down with a player indicating they want to play a specific character class or have some specific ability and adjusting the random generation to help them achieve that.

      Delete
    3. That's not a bad approach. Someone mentioned starting with a set score at some attributes and then rolling fewer or smaller dice to add.

      Delete
  11. FWIW I prefer a set array of scores over point buy or fiddly swap/extra roll system.
    Perhaps the value of the scores are set, but a system to randomize where they (or some of them) get assigned?
    I guess that depends on how much you want to have players play the exact character they want, or deal with the fickle fates.
    (maybe randomly determine where the highest and lowest go, and the player gets to pick the rest.) But that may be fiddly tooo?

    ReplyDelete
    Replies
    1. As discussed on an earlier attributes post, point buy or array doesn't always work. RuneQuest has some "uber" attributes like INT where I would expect many people to put their highest array value in. I'd put my 2nd highest in SIZ if I wanted to be a good warrior (CON and STR can be increased to highest of STR, CON, SIZ). Not sure how to distribute the rest (and does anyone have a 7 attribute standard array handy?).

      I also disliked the D&D 3.x standard array for having almost every value in it odd values when bonuses break at even values.

      Delete
  12. The Amber Diceless system expected and accounted for player rivalry. The four abilities were put up for auction but a player didn’t have to spend all of their points then, which could also be used later for powers like the Pattern. The auction established discrete levels of the abilities. Afterwards players could spend leftover points to raise their bid to a higher bid (but nothing in between). The character with the highest bid was still considered marginally superior even to those who raised their scores to match.

    Being diceless, Amber didn’t have any real mechanics underlying this. But I could see another system establishing a penalty of -1 for the lowest score and a bonus of +1 for the highest (or maybe +2 with +1 for the second highest, and a prohibition on matching the highest during PC creation).

    This also makes the meaning of the ability scores revolve around the PCs to an extent that worked for Amber but probably not a lot of other settings.

    ReplyDelete
  13. One that we've often had fun with is for the GM to roll a random set, each player then takes the numbers and assigns them to the attributes of their choice. Almost always has a negative number or two, often a decent score for at least one stat, and quite a few that are around average, and no-one feels hard done by because another player rolled a bunch of high teens while they ended up with everything below 9.

    Then again, I've had a whale of a time sometimes with characters who've had a highest stat of 8, even despite given the chance to reroll them.

    ReplyDelete
  14. "Swords & Six-Siders has a rather ingenious methods. Stats are from 1-6. You roll three dice and your stats are the numbers on top AND on the bottom of the dice. So if you roll three sixes, you get 6,6,6,1,1,1 arrange as you like. If your roll three 3's, you get 3,3,3,4,4,4, etc. I'm not sure how you could get something like that to work with 3d6, but I like the idea.

    ReplyDelete